Here’s the web-app number before the sales call: a browser-based application in 2026 runs $20K to $70K for a simple internal tool, $80K to $180K for a medium build, and $200K to $500K or more for a complex one. Put a finer point on it by what the app actually is: a CRUD internal tool sits at the bottom, a customer portal or standard SaaS in the middle, a multi-tenant SaaS platform at the top. Where you land has almost nothing to do with how many screens you sketch.
It has everything to do with three things: how many user roles see different data, how many other systems the app talks to, and whether it’s multi-tenant. A one-role dashboard and a billing platform can both look polished in a mockup and cost five times apart. The mockup is the cheap part.
We’re gmware, a software development firm headquartered in Austin, TX with engineering centers in Bangalore and Mohali, India, and web and product builds are a core line of work for us. This guide covers browser-based web apps only, not native mobile. Below: cost by app type, how the stack moves the number, the build phases, the hosting and maintenance bill quotes skip, and when a no-code platform beats a custom build outright.
2026 web app cost at a glance
What a web application costs in 2026, by type
The complexity bands give you a range, but “medium complexity” is a lousy thing to budget against. So here’s the same spend re-cut by what the app does, which is the lens a founder actually thinks in. The complexity ranges behind these are well documented: simple $20K to $70K, medium $80K to $180K, complex $200K to $500K+, and other 2026 estimates land in the same neighborhood with medium builds at $60K to $150K and complex at $150K to $350K+.
| Web app type | What it is | Roles / tenancy | Typical cost | Timeline |
|---|---|---|---|---|
| CRUD internal tool | Forms over a database, light reporting, one team | 1 to 2 roles, single tenant | $20K to $50K | 4 to 10 weeks |
| Customer portal | Login, account view, documents, support, light self-service | 2 to 3 roles, single tenant | $60K to $120K | 3 to 5 months |
| Standard SaaS | Signup, billing, an admin view, a few integrations | 2 to 4 roles, single or multi-tenant | $80K to $180K | 4 to 7 months |
| Multi-tenant SaaS at scale | Tenant isolation, heavy integrations, compliance, high volume | Many roles, multi-tenant | $200K to $500K+ | 9 to 15+ months |
This is the original artifact, so read the columns that move money: roles and tenancy. A customer portal and a CRUD tool might share a screen count, but the portal authenticates real customers, shows each one only their own data, and handles support, which is why it sits two or three times higher. SaaS platforms specifically tend to run $50K to $200K+, and the jump from “standard SaaS” to “multi-tenant at scale” is the single biggest step on the table. Tenant isolation is not a feature you sprinkle on later. It’s an architecture decision that touches the database, the auth layer, and every query.
Cost by web app type, on one scale
Roles and integrations cost more than screens
Because a screen is a screen. A button that opens a form is cheap, and it stays cheap whether you have ten of them or thirty. What’s expensive is everything behind the button, and that’s where web-app quotes quietly inflate.
Each user role is not just a login. It’s a permission system, a separate view of the data, and a separate set of paths to test. Two roles isn’t double one role, but it’s a long way from equal. Three roles with a tenant boundary between customers is where the cost curve bends hard.
Integrations are the other multiplier. Every external system the app talks to, a payment processor, a CRM, an email service, an accounting tool, needs its own connection code, its own handling for when that system is slow or down, and its own tests. We’ve watched the integration line surprise people more than any other number on a quote. The first web app we scoped this way, we under-counted the data-mapping work that happens before the integration even runs, and we don’t make that mistake anymore. If the app has to talk to four systems, four is doing most of the pricing.
Compliance gets its own mention. If the app handles health or financial data, you’re designing audit logging, access controls, and data-region rules from day one. Bolting those on after launch costs more than building them in. Every time.
How the technology stack moves the number
Mostly through speed and hiring, not license fees, which is the part buyers get backwards. The frameworks that dominate web builds are open source and free: open-source frameworks like Laravel or Ruby on Rails carry no license cost, and the same goes for React, Django, Node, and the rest of the usual front-end and back-end set. So the stack rarely shows up as a line item you can point at.
Where it shows up is the rate and the timeline. A stack your team already knows ships faster, because nobody’s learning on your budget. A familiar stack markedly reduces the development timeline, and timeline is cost. The expensive stack decision is the opposite move: picking something niche that few engineers know, which narrows your hiring pool, raises the rate, and slows everything down. We’ve inherited a few of those. A clever, obscure stack chosen by one departed developer is a tax the next team pays for years.
The honest version: for most web apps, the stack is close to a non-decision on price. Pick mainstream tools, pick what the team building it is fluent in, and spend your attention on scope instead. That’s where the real money is.
What the build phases look like
A web app ships in phases, and the budget follows the phases rather than landing as one lump. The shares are fairly consistent across 2026 estimates: discovery and planning around 10% of the budget, development consuming 60% to 70%, with design, testing, and deployment splitting the rest. Here’s a representative breakdown for a mid-tier portal or SaaS build.
| Phase | Share of budget | Work |
|---|---|---|
| Discovery & planning | ~10% | Scope, data model, role map, integration inventory |
| UI/UX design | 10% to 15% | Flows, screens, design system |
| Development | 60% to 70% | Front end, back end, database, integrations |
| Testing & QA | 10% to 15% | Functional, security, cross-browser, load |
| Deployment | small | Infrastructure, CI/CD, launch |
Two things worth flagging. Discovery is a real phase, not a free pre-sales gift, and the cheapest hour on this whole table is the one where you decide what not to build. And testing is not where you save money. A web app that skips the security and cross-browser pass is the one that breaks in front of a customer. We itemize the same phase logic for adjacent builds in our SaaS MVP cost breakdown, where the plumbing-versus-screens split is just as stark.
What it costs to host and maintain a web app
Software is never finished, and the run cost is the line most teams forget until the invoice lands. It has two parts: maintenance and hosting.
Maintenance runs 15% to 25% of the original build cost per year, a benchmark that’s held across decades of enterprise software data. On a $100K build that’s $15K to $25K annually, and it’s not padding. It covers security patches, dependency upgrades, browser changes, and the small fixes that keep a live system live. Skip it and the cost doesn’t vanish, it defers into an emergency repair at emergency rates.
Hosting is separate and scales with traffic. Early-stage apps run about $15 to $70 a month, growth-stage apps $100 to $400, and enterprise platforms with heavier traffic and compliance needs $850 to $4,300 a month. The vector most people fear, a runaway cloud bill, is real only at scale and only if nobody’s watching it. For a portal with a few thousand users, hosting is a rounding error next to the maintenance retainer.
Monthly hosting by stage
How a US-managed, India-delivery model changes the math
The rate spread between regions is the biggest single lever on a web-app quote, and it’s enormous. North American developers bill $120 to $250+ an hour, while teams in Asia quote $20 to $45, with senior engineers in the Asia Pacific region in the $60 to $130+ range. That spread is the whole reason offshore delivery exists.
But the quoted rate is not the delivered cost, and the cheapest rate is rarely the cheapest finished app. Pure rock-bottom offshore tends to hide a rework multiplier: developers rotate through the codebase, scope drifts, and you pay for the same learning curve more than once. The model we run keeps the rate low without the accountability gap: an Austin-side architect who owns scope and code review, engineering from Bangalore and Mohali at India economics, three to four hours of daily overlap, and IP assigned to you under a US master services agreement.
That’s the same structure, and the same honesty about its trade-offs, that we lay out in our custom software cost guide for small businesses. The rate gets you the engineers. The overlap and the review gates get you software that works.
When a no-code platform beats a custom build
More often than agencies admit, including for problems we’d otherwise quote. If the app is an internal CRUD tool over data you already have, with a handful of users and no public-facing or compliance demands, a low-code platform will beat a custom build on every axis that matters.
Retool is the clearest example. Its Free plan covers up to 5 users, the Team plan is $10 per builder a month plus $5 per internal user, and the Business plan is $50 per builder with SSO and audit logging. For an ops dashboard or an admin panel that five people touch, that’s a few hundred dollars a year against a $30K-plus custom build. The math isn’t close.
Then it flips back to build. The honest fork:
- Build no-code (Retool, similar) when it’s an internal tool, the users are your own team, the data lives in systems you already connect to, and nobody outside the company logs in.
- Build custom when the app is your product and your differentiation, when it must scale to many tenants or many thousands of users, when it’s public-facing with your brand on it, or when compliance rules out a shared third-party platform handling your data.
The trap is using a no-code platform past its ceiling, then paying twice when you rebuild custom for scale. We’ve cleaned up that exact migration. The other trap is the reverse: spending $40K building a custom internal tool that Retool would have run for the price of a lunch. Pick the side of the fork your actual requirements land on, not the one that flatters the project.
No-code or custom: which side are you on
What we’d recommend
For most teams the right first web app is smaller than the one they walk in describing. We’d rather ship a tight customer portal that proves one workflow than a multi-tenant everything-platform that takes a year to find out the workflow was wrong. Our product development engagements open with a fixed-scope discovery, so the number and the cut list are real before you sign, not invented in the third meeting.
And sometimes we say don’t build. If the app is an internal tool five people will use, we’ll point you at a no-code platform and save you the quote. If you’re still weighing what a full custom build costs once an MVP proves out, or how to vet the shop you hire, our guides on SaaS MVP cost and choosing a software development company carry the decision forward. (Building for phones too? The mobile app development cost math runs separately, since native is a different beast.)
Tell us what you’re trying to build and we’ll give you a straight answer on scope, cost, and timeline within 48 hours, with a real range in the first conversation. Reach out.